Back to cybriq.io The Cross Family is part of CybrIQ

Six products. Four jobs.

Your rooms run on a network nobody documented.

Two of these tools tell you what is on that network. One checks the documentation you already have and tells you which parts of it are still true. Two tell you whether the room is going to work, and why it did not. The sixth steps back and runs the building all of it sits in. You buy, install and use each one on its own, but they share a way of working: they only read, and every answer arrives with the evidence behind it.

Know what is there.

the record

Nobody signs off on an upgrade to a building they have to guess about. One of these gives you the answer this week and leaves; the other stays and runs the estate off it.

One is a scanner. The other is a platform.

They are not two sizes of the same thing. CrossScan produces an inventory. CrossConnect runs your estate off an inventory. If you need a list of what is in a building, CrossConnect is far more product than you need, and you will pay for all of it.

  CrossScan CrossConnect
What it is A scanner.You point it, it reads, it hands you the answer and gets out of the way. A platform.It holds the estate and everything you do to the estate runs through it.
What it produces An inventory.Every port, the device on it, the room, the confidence, the evidence. Plus low-level health: PoE draw, link errors, per-room AV readiness. An operating picture.It takes the inventory as a starting point and then goes further: what is misconfigured, what breaks if this link dies, where capacity runs out, what changed at 2am, and why the application is slow.
How far it reaches A site.One seed switch, out to the switches behind it. The whole estate.Every site, continuously, with routing, addressing, capacity, compliance and change history on top.
How long it lives An afternoon.Run it, export the report, close it. Run it again next quarter and it tells you what moved. Permanently.It keeps discovering on its own, so the picture is still true in eight months.
Where it runs Your laptop.A desktop app on Windows or macOS. Nothing to host, nothing to stand up. Your server.One command, inside your walls. It brings its own database.
Do you need both? Usually no. Start with the job you have.If the job is "tell me what is in this building," a scanner does it and a platform is overkill. If the job is "run this estate without being wrong," a scan will not keep you honest past the first cable move.

Know what you were told.

the paperwork

Before any of that, there is a folder of drawings, spreadsheets and configs that somebody already paid for. It describes your network. Nobody can tell you which parts of it are still true, so nobody acts on any of it.

CrossTrust

The fact-check. Point it at a folder and it reads your existing documentation, then shows you where your own records contradict each other, with both sides cited to the line. It reads 11 configuration languages, runs 71 settings checks, and when two sources are evenly matched it refuses to pick and tells you the command that settles it.

"We have documentation. Nobody trusts a word of it."

See CrossTrust →

It never touches your network. Everything CrossTrust knows came out of a file you already had, which is why it needs no credentials, no access and no change window. That also fixes its limit in place: when it tells you a switch is set a certain way, it means a file in your folder says so, and it dates that file on screen so you can weigh it yourself.

If you want the network read directly rather than the paperwork about it, that is CrossScan above, or CrossConnect.

Know what the building costs.

the bill

The rooms sit inside a building, and somebody has to run it: the plant, the water, whatever is on tonight and the electricity bill it all lands on. A concert hall, a hospital, an office tower, a convention center. That job is usually done from memory, by one person, and it is the largest number in the estate.

CrossVenue

The operations platform. It reads a venue and answers the calls somebody has to make today, each with the arithmetic under it. Then it replays the last 30 recorded nights against 7,140 settings to say what they should have been run at, and prints the cheaper answer it refused because a seat would have been too warm.

"Finance asked what a full house costs to run and I could not tell them."

See CrossVenue →

It never writes a setpoint. This is the one product here that could plausibly be asked to drive something, and it holds the same line as the other five: it reads the controls, works out what to do, and hands it to a person. Every source it reads is asked for an account that cannot write.

It runs one site at a time. If the question is about a network across many buildings, that is CrossConnect above.

How they fit together

Six questions. One rule: show the evidence or say nothing.

Buy one and it works alone. What they share is a discipline. All six read and never write. All six show you the evidence behind every number they put on a screen.

The same three rules, in all six

Read-only None of them writes to a device: no SNMP set, no config push, no traffic mirroring. That is one decision to trust, and you make it once.
Evidence, or silence Every value carries where it came from and how it was known. Confirmed beats imported beats inferred, and the product says which one you are looking at.
Missing data lowers the score A gap caps the verdict and names what would lift the cap. None of them rounds an unknown up to a green light.

The thing all six do

When one of these tools is not sure, it lowers the score and tells you why.

Any tool can sound confident. The hard part is being uncertain in a way you can act on. So a gap in the evidence is never rounded away or quietly filled with the likeliest answer: it comes off the score, it gets named, and the one thing that would close it is named beside it. Here is what that costs, in each product, as a real number.

no answer

What CrossTrust returns when two documents disagree and neither outweighs the other. It names the command that settles it instead.

60

The highest a CrossCheck verdict can reach when the firmware came from your memory. A Go starts at 75.

74

The highest a CrossRoom verdict can reach when a layer was not instrumented. Good starts at 80.

14

Devices in the CrossScan demo it declined to name, out of 43. All fourteen still on the map, in grey.

unknown

What CrossConnect returns for a service it has no telemetry for. Never green.

48

Findings CrossVenue puts no price on, out of 60, because the rate they would rest on has not been supplied.

  • CrossTrust: it will not break a tie it cannot break A value is only ever named as current on a strict win: it has to be legal, outrank every source that disagrees, and be no older than any of them. Two evenly matched documents get no winner and a named next step instead, because the satisfying answer and the true one are not reliably the same answer.
  • CrossScan: unknown is a finding In the scan on its own page, fourteen of forty-three devices stayed quiet and it could not identify them. They are all still on the map, in grey, listed by port, with the little it does know about each. Leaving them off would be the easiest lie in the product.
  • CrossCheck: you cannot type your way to a Go A firmware version you typed from memory is trusted at 60%, and the whole verdict is multiplied by it. A perfect change on a model you typed still lands on 60, which is a Caution. Read the device instead and the multiplier disappears.
  • CrossRoom: a half-seen room cannot look healthy Miss a layer that mattered and the verdict is capped at 74. Good starts at 80. The arithmetic makes it impossible for an under-instrumented room to present as fine, and the report names the integration that would lift the cap.
  • CrossConnect: it will say it does not know A service with no telemetry comes back unknown rather than green. A service that is healthy inside your walls says, in writing, that the far end is unobserved because no flow source is wired. It never claims to have checked something it did not check.
  • CrossVenue: it counts its live feeds instead of claiming them The top of every screen carries how many of the nine sources are really connected, and in the demo that number is zero. The money screen grades itself Indicative while the rates it uses are assumed, and forty-eight of its sixty findings carry no figure at all rather than a plausible one.

The limits

The same honesty, drawn as a line none of them will cross.

A score that lowers itself is one half of it. The other half is a set of things these tools simply will not do, no matter how convenient it would be to do them.

  • Read-only. No SNMP set, no config push, no traffic mirroring. They read what the network already knows about itself and hand you the change to make.
  • Everything runs on your own machine or your own server. Each product has a feature or two that reaches outside your walls, the AI assist among them. Every one is off until you switch it on, and shows you what it would send. Leave them all off and the products still work.
  • A red light is where the work starts. These are diagnostic tools, and a tool that hands you an alarm and stops has left you where it found you.
  • None of them is finished on day one. Each starts with what it can read on its own and gets sharper as you connect the next source. The first verdict is a to-do list. The tenth is one you would stake a change window on.

Try one

Nothing you click will change anything. Neither does the product.

Six demos, one per product. Nothing to fill in and nothing to install. Each one takes about a minute and ends on a real screen, with a line telling you what to look at.